| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 939, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.003194888178913738, | |
| "grad_norm": 22.587236912668185, | |
| "learning_rate": 5.3191489361702125e-09, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -104.0, | |
| "logps/rejected": -105.5, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03194888178913738, | |
| "grad_norm": 20.504214733036054, | |
| "learning_rate": 5.3191489361702123e-08, | |
| "logits/chosen": -3.78125, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -112.5, | |
| "logps/rejected": -127.5, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.0416666679084301, | |
| "rewards/chosen": -0.0045166015625, | |
| "rewards/margins": -0.0020904541015625, | |
| "rewards/rejected": -0.0024261474609375, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06389776357827476, | |
| "grad_norm": 26.543460167609325, | |
| "learning_rate": 1.0638297872340425e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.921875, | |
| "logps/chosen": -131.0, | |
| "logps/rejected": -141.0, | |
| "loss": 0.6905, | |
| "rewards/accuracies": 0.012500000186264515, | |
| "rewards/chosen": 0.000308990478515625, | |
| "rewards/margins": -0.00093841552734375, | |
| "rewards/rejected": 0.001251220703125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.09584664536741214, | |
| "grad_norm": 90.02360712811911, | |
| "learning_rate": 1.5957446808510638e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -126.5, | |
| "logps/rejected": -132.0, | |
| "loss": 0.6906, | |
| "rewards/accuracies": 0.10000000149011612, | |
| "rewards/chosen": 0.0128173828125, | |
| "rewards/margins": 0.002197265625, | |
| "rewards/rejected": 0.0106201171875, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.12779552715654952, | |
| "grad_norm": 17.123125909083377, | |
| "learning_rate": 2.127659574468085e-07, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -135.0, | |
| "logps/rejected": -133.0, | |
| "loss": 0.6896, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": 0.034423828125, | |
| "rewards/margins": 0.00689697265625, | |
| "rewards/rejected": 0.027587890625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.1597444089456869, | |
| "grad_norm": 31.72995333133769, | |
| "learning_rate": 2.659574468085106e-07, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -116.0, | |
| "logps/rejected": -130.0, | |
| "loss": 0.6862, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.051513671875, | |
| "rewards/margins": 0.015625, | |
| "rewards/rejected": 0.035888671875, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19169329073482427, | |
| "grad_norm": 27.598228962388703, | |
| "learning_rate": 3.1914893617021275e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -123.0, | |
| "logps/rejected": -124.0, | |
| "loss": 0.6864, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.1298828125, | |
| "rewards/margins": 0.0147705078125, | |
| "rewards/rejected": 0.115234375, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22364217252396165, | |
| "grad_norm": 24.68090794587246, | |
| "learning_rate": 3.7234042553191484e-07, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -114.0, | |
| "logps/rejected": -120.5, | |
| "loss": 0.6794, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.2099609375, | |
| "rewards/margins": 0.0250244140625, | |
| "rewards/rejected": 0.185546875, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.25559105431309903, | |
| "grad_norm": 32.31032873231678, | |
| "learning_rate": 4.25531914893617e-07, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -128.0, | |
| "logps/rejected": -141.0, | |
| "loss": 0.6746, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.359375, | |
| "rewards/margins": 0.042724609375, | |
| "rewards/rejected": 0.31640625, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.28753993610223644, | |
| "grad_norm": 25.62788513297419, | |
| "learning_rate": 4.787234042553192e-07, | |
| "logits/chosen": -3.921875, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -125.5, | |
| "logps/rejected": -134.0, | |
| "loss": 0.6756, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.33203125, | |
| "rewards/margins": 0.07373046875, | |
| "rewards/rejected": 0.259765625, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.3194888178913738, | |
| "grad_norm": 34.430976745441534, | |
| "learning_rate": 4.964497041420119e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -115.0, | |
| "logps/rejected": -116.0, | |
| "loss": 0.6704, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.30078125, | |
| "rewards/margins": 0.027587890625, | |
| "rewards/rejected": 0.2734375, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3514376996805112, | |
| "grad_norm": 19.726440993478906, | |
| "learning_rate": 4.905325443786982e-07, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -123.0, | |
| "logps/rejected": -132.0, | |
| "loss": 0.6666, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.361328125, | |
| "rewards/margins": 0.126953125, | |
| "rewards/rejected": 0.2353515625, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38338658146964855, | |
| "grad_norm": 22.606814351335732, | |
| "learning_rate": 4.846153846153846e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -116.5, | |
| "logps/rejected": -120.5, | |
| "loss": 0.6611, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": 0.36328125, | |
| "rewards/margins": 0.07373046875, | |
| "rewards/rejected": 0.2890625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.41533546325878595, | |
| "grad_norm": 30.108481813541545, | |
| "learning_rate": 4.78698224852071e-07, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -108.5, | |
| "logps/rejected": -133.0, | |
| "loss": 0.6624, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": 0.462890625, | |
| "rewards/margins": 0.2060546875, | |
| "rewards/rejected": 0.255859375, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4472843450479233, | |
| "grad_norm": 42.69476922343966, | |
| "learning_rate": 4.727810650887574e-07, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -122.0, | |
| "logps/rejected": -131.0, | |
| "loss": 0.6584, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.07666015625, | |
| "rewards/margins": 0.09033203125, | |
| "rewards/rejected": -0.013427734375, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.4792332268370607, | |
| "grad_norm": 29.2950267437935, | |
| "learning_rate": 4.668639053254438e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -122.0, | |
| "logps/rejected": -129.0, | |
| "loss": 0.6454, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.0849609375, | |
| "rewards/margins": 0.265625, | |
| "rewards/rejected": -0.1806640625, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5111821086261981, | |
| "grad_norm": 29.18218174433763, | |
| "learning_rate": 4.6094674556213014e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -126.0, | |
| "logps/rejected": -135.0, | |
| "loss": 0.6396, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -0.1611328125, | |
| "rewards/margins": 0.2197265625, | |
| "rewards/rejected": -0.380859375, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5431309904153354, | |
| "grad_norm": 24.071324961932216, | |
| "learning_rate": 4.5502958579881655e-07, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.921875, | |
| "logps/chosen": -102.5, | |
| "logps/rejected": -104.0, | |
| "loss": 0.6336, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.31640625, | |
| "rewards/margins": 0.19140625, | |
| "rewards/rejected": 0.125, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5750798722044729, | |
| "grad_norm": 15.807029915813834, | |
| "learning_rate": 4.491124260355029e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -115.5, | |
| "logps/rejected": -122.5, | |
| "loss": 0.63, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.45703125, | |
| "rewards/margins": 0.3359375, | |
| "rewards/rejected": -0.79296875, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.6070287539936102, | |
| "grad_norm": 25.188943189422474, | |
| "learning_rate": 4.4319526627218936e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -123.5, | |
| "logps/rejected": -137.0, | |
| "loss": 0.6489, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -0.81640625, | |
| "rewards/margins": 0.1787109375, | |
| "rewards/rejected": -0.99609375, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6389776357827476, | |
| "grad_norm": 27.085036945419034, | |
| "learning_rate": 4.372781065088757e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -116.0, | |
| "logps/rejected": -123.0, | |
| "loss": 0.6206, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -0.357421875, | |
| "rewards/margins": 0.2431640625, | |
| "rewards/rejected": -0.6015625, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.670926517571885, | |
| "grad_norm": 59.95656852986459, | |
| "learning_rate": 4.313609467455621e-07, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -120.0, | |
| "logps/rejected": -122.0, | |
| "loss": 0.6614, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -0.2216796875, | |
| "rewards/margins": 0.1650390625, | |
| "rewards/rejected": -0.38671875, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7028753993610224, | |
| "grad_norm": 10.470332523245947, | |
| "learning_rate": 4.2544378698224847e-07, | |
| "logits/chosen": -3.96875, | |
| "logits/rejected": -3.984375, | |
| "logps/chosen": -122.5, | |
| "logps/rejected": -133.0, | |
| "loss": 0.6319, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.0751953125, | |
| "rewards/margins": 0.197265625, | |
| "rewards/rejected": -0.1220703125, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.7348242811501597, | |
| "grad_norm": 29.016293729576255, | |
| "learning_rate": 4.195266272189349e-07, | |
| "logits/chosen": -3.984375, | |
| "logits/rejected": -3.984375, | |
| "logps/chosen": -146.0, | |
| "logps/rejected": -135.0, | |
| "loss": 0.6889, | |
| "rewards/accuracies": 0.15000000596046448, | |
| "rewards/chosen": -0.05224609375, | |
| "rewards/margins": 0.10205078125, | |
| "rewards/rejected": -0.154296875, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.7667731629392971, | |
| "grad_norm": 29.554961998726093, | |
| "learning_rate": 4.1360946745562133e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.9375, | |
| "logps/chosen": -112.5, | |
| "logps/rejected": -125.0, | |
| "loss": 0.6383, | |
| "rewards/accuracies": 0.13750000298023224, | |
| "rewards/chosen": 0.00262451171875, | |
| "rewards/margins": 0.1708984375, | |
| "rewards/rejected": -0.16796875, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7987220447284346, | |
| "grad_norm": 30.050484609489388, | |
| "learning_rate": 4.076923076923077e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.984375, | |
| "logps/chosen": -109.0, | |
| "logps/rejected": -119.5, | |
| "loss": 0.6651, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": 0.35546875, | |
| "rewards/margins": 0.310546875, | |
| "rewards/rejected": 0.046630859375, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.8306709265175719, | |
| "grad_norm": 50.35008935486711, | |
| "learning_rate": 4.017751479289941e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -3.953125, | |
| "logps/chosen": -123.5, | |
| "logps/rejected": -135.0, | |
| "loss": 0.6779, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.455078125, | |
| "rewards/margins": 0.1943359375, | |
| "rewards/rejected": 0.26171875, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.8626198083067093, | |
| "grad_norm": 33.675183816708014, | |
| "learning_rate": 3.9585798816568044e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.96875, | |
| "logps/chosen": -104.5, | |
| "logps/rejected": -109.0, | |
| "loss": 0.6701, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.6171875, | |
| "rewards/margins": 0.1103515625, | |
| "rewards/rejected": 0.5078125, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.8945686900958466, | |
| "grad_norm": 21.039116456131303, | |
| "learning_rate": 3.8994082840236685e-07, | |
| "logits/chosen": -3.921875, | |
| "logits/rejected": -3.921875, | |
| "logps/chosen": -110.0, | |
| "logps/rejected": -112.5, | |
| "loss": 0.6536, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.5546875, | |
| "rewards/margins": 0.0791015625, | |
| "rewards/rejected": 0.474609375, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.9265175718849841, | |
| "grad_norm": 29.646688523670253, | |
| "learning_rate": 3.840236686390532e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.953125, | |
| "logps/chosen": -125.5, | |
| "logps/rejected": -135.0, | |
| "loss": 0.6689, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -0.06982421875, | |
| "rewards/margins": 0.228515625, | |
| "rewards/rejected": -0.296875, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.9584664536741214, | |
| "grad_norm": 24.41030402543526, | |
| "learning_rate": 3.7810650887573966e-07, | |
| "logits/chosen": -3.984375, | |
| "logits/rejected": -4.0, | |
| "logps/chosen": -131.0, | |
| "logps/rejected": -136.0, | |
| "loss": 0.6724, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": -0.2734375, | |
| "rewards/margins": 0.08837890625, | |
| "rewards/rejected": -0.36328125, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.9904153354632588, | |
| "grad_norm": 19.87954624614089, | |
| "learning_rate": 3.72189349112426e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.9375, | |
| "logps/chosen": -133.0, | |
| "logps/rejected": -137.0, | |
| "loss": 0.6607, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -0.05224609375, | |
| "rewards/margins": 0.1044921875, | |
| "rewards/rejected": -0.15625, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -3.90625, | |
| "eval_logits/rejected": -3.921875, | |
| "eval_logps/chosen": -104.5, | |
| "eval_logps/rejected": -109.5, | |
| "eval_loss": 0.6517187356948853, | |
| "eval_rewards/accuracies": 0.1964285671710968, | |
| "eval_rewards/chosen": -0.1142578125, | |
| "eval_rewards/margins": 0.1513671875, | |
| "eval_rewards/rejected": -0.265625, | |
| "eval_runtime": 12.1044, | |
| "eval_samples_per_second": 16.523, | |
| "eval_steps_per_second": 0.578, | |
| "step": 313 | |
| }, | |
| { | |
| "epoch": 1.0223642172523961, | |
| "grad_norm": 10.097984976496987, | |
| "learning_rate": 3.662721893491124e-07, | |
| "logits/chosen": -3.984375, | |
| "logits/rejected": -3.984375, | |
| "logps/chosen": -135.0, | |
| "logps/rejected": -144.0, | |
| "loss": 0.5903, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": 0.1611328125, | |
| "rewards/margins": 0.86328125, | |
| "rewards/rejected": -0.703125, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.0543130990415335, | |
| "grad_norm": 1.1859208612609746, | |
| "learning_rate": 3.603550295857988e-07, | |
| "logits/chosen": -4.25, | |
| "logits/rejected": -4.25, | |
| "logps/chosen": -130.0, | |
| "logps/rejected": -151.0, | |
| "loss": 0.5324, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -1.359375, | |
| "rewards/margins": 1.1484375, | |
| "rewards/rejected": -2.515625, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.0862619808306708, | |
| "grad_norm": 2.418489775814248, | |
| "learning_rate": 3.544378698224852e-07, | |
| "logits/chosen": -4.46875, | |
| "logits/rejected": -4.5, | |
| "logps/chosen": -139.0, | |
| "logps/rejected": -154.0, | |
| "loss": 0.5059, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -1.703125, | |
| "rewards/margins": 1.3671875, | |
| "rewards/rejected": -3.0625, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.1182108626198084, | |
| "grad_norm": 14.519066358162673, | |
| "learning_rate": 3.485207100591716e-07, | |
| "logits/chosen": -4.15625, | |
| "logits/rejected": -4.15625, | |
| "logps/chosen": -133.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.4855, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -1.6796875, | |
| "rewards/margins": 1.9609375, | |
| "rewards/rejected": -3.640625, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.1501597444089458, | |
| "grad_norm": 10.64966368268835, | |
| "learning_rate": 3.42603550295858e-07, | |
| "logits/chosen": -4.125, | |
| "logits/rejected": -4.1875, | |
| "logps/chosen": -148.0, | |
| "logps/rejected": -186.0, | |
| "loss": 0.535, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -2.265625, | |
| "rewards/margins": 1.671875, | |
| "rewards/rejected": -3.9375, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.182108626198083, | |
| "grad_norm": 16.274756270248496, | |
| "learning_rate": 3.366863905325444e-07, | |
| "logits/chosen": -4.1875, | |
| "logits/rejected": -4.21875, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -170.0, | |
| "loss": 0.5207, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -2.046875, | |
| "rewards/margins": 1.5, | |
| "rewards/rejected": -3.546875, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.2140575079872205, | |
| "grad_norm": 2.316873389478435, | |
| "learning_rate": 3.3076923076923075e-07, | |
| "logits/chosen": -4.09375, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -130.0, | |
| "logps/rejected": -145.0, | |
| "loss": 0.5208, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": -1.8359375, | |
| "rewards/margins": 1.2890625, | |
| "rewards/rejected": -3.125, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.2460063897763578, | |
| "grad_norm": 13.428350268463888, | |
| "learning_rate": 3.2485207100591716e-07, | |
| "logits/chosen": -4.09375, | |
| "logits/rejected": -4.125, | |
| "logps/chosen": -144.0, | |
| "logps/rejected": -167.0, | |
| "loss": 0.5454, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -1.71875, | |
| "rewards/margins": 1.4453125, | |
| "rewards/rejected": -3.15625, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.2779552715654952, | |
| "grad_norm": 1.0689515086010029, | |
| "learning_rate": 3.189349112426035e-07, | |
| "logits/chosen": -4.09375, | |
| "logits/rejected": -4.09375, | |
| "logps/chosen": -159.0, | |
| "logps/rejected": -182.0, | |
| "loss": 0.5154, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -2.34375, | |
| "rewards/margins": 1.8828125, | |
| "rewards/rejected": -4.21875, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3099041533546325, | |
| "grad_norm": 23.626902495164178, | |
| "learning_rate": 3.130177514792899e-07, | |
| "logits/chosen": -4.125, | |
| "logits/rejected": -4.09375, | |
| "logps/chosen": -151.0, | |
| "logps/rejected": -172.0, | |
| "loss": 0.4904, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -2.453125, | |
| "rewards/margins": 1.7578125, | |
| "rewards/rejected": -4.1875, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.34185303514377, | |
| "grad_norm": 11.414992983326627, | |
| "learning_rate": 3.071005917159763e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -4.125, | |
| "logps/chosen": -151.0, | |
| "logps/rejected": -184.0, | |
| "loss": 0.4783, | |
| "rewards/accuracies": 0.38749998807907104, | |
| "rewards/chosen": -2.1875, | |
| "rewards/margins": 2.390625, | |
| "rewards/rejected": -4.5625, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.3738019169329074, | |
| "grad_norm": 0.7721195435055699, | |
| "learning_rate": 3.011834319526627e-07, | |
| "logits/chosen": -4.0625, | |
| "logits/rejected": -4.125, | |
| "logps/chosen": -142.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.5043, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -2.796875, | |
| "rewards/margins": 1.796875, | |
| "rewards/rejected": -4.59375, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.4057507987220448, | |
| "grad_norm": 31.030290011978142, | |
| "learning_rate": 2.952662721893491e-07, | |
| "logits/chosen": -4.0625, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -170.0, | |
| "loss": 0.5087, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -2.953125, | |
| "rewards/margins": 1.5234375, | |
| "rewards/rejected": -4.46875, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.4376996805111821, | |
| "grad_norm": 56.770434850318985, | |
| "learning_rate": 2.893491124260355e-07, | |
| "logits/chosen": -4.0625, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -145.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.5115, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -2.3125, | |
| "rewards/margins": 2.03125, | |
| "rewards/rejected": -4.34375, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.4696485623003195, | |
| "grad_norm": 4.7100434984872575, | |
| "learning_rate": 2.834319526627219e-07, | |
| "logits/chosen": -4.09375, | |
| "logits/rejected": -4.09375, | |
| "logps/chosen": -161.0, | |
| "logps/rejected": -177.0, | |
| "loss": 0.4992, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -3.484375, | |
| "rewards/margins": 1.5546875, | |
| "rewards/rejected": -5.03125, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.5015974440894568, | |
| "grad_norm": 15.230105133264576, | |
| "learning_rate": 2.7751479289940824e-07, | |
| "logits/chosen": -4.1875, | |
| "logits/rejected": -4.15625, | |
| "logps/chosen": -148.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.5108, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -3.625, | |
| "rewards/margins": 2.140625, | |
| "rewards/rejected": -5.78125, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.5335463258785942, | |
| "grad_norm": 10.678316470819308, | |
| "learning_rate": 2.715976331360947e-07, | |
| "logits/chosen": -4.15625, | |
| "logits/rejected": -4.15625, | |
| "logps/chosen": -151.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.521, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -3.796875, | |
| "rewards/margins": 1.4765625, | |
| "rewards/rejected": -5.28125, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.5654952076677318, | |
| "grad_norm": 8.439588364665873, | |
| "learning_rate": 2.6568047337278106e-07, | |
| "logits/chosen": -4.1875, | |
| "logits/rejected": -4.21875, | |
| "logps/chosen": -152.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.5203, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -3.296875, | |
| "rewards/margins": 2.28125, | |
| "rewards/rejected": -5.5625, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.5974440894568689, | |
| "grad_norm": 7.316694492795425, | |
| "learning_rate": 2.5976331360946746e-07, | |
| "logits/chosen": -4.125, | |
| "logits/rejected": -4.15625, | |
| "logps/chosen": -172.0, | |
| "logps/rejected": -205.0, | |
| "loss": 0.5227, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -3.28125, | |
| "rewards/margins": 1.984375, | |
| "rewards/rejected": -5.28125, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.6293929712460065, | |
| "grad_norm": 8.243916218487573, | |
| "learning_rate": 2.538461538461538e-07, | |
| "logits/chosen": -4.125, | |
| "logits/rejected": -4.125, | |
| "logps/chosen": -167.0, | |
| "logps/rejected": -188.0, | |
| "loss": 0.5044, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -3.453125, | |
| "rewards/margins": 2.125, | |
| "rewards/rejected": -5.5625, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.6613418530351438, | |
| "grad_norm": 7.359358015998256, | |
| "learning_rate": 2.479289940828402e-07, | |
| "logits/chosen": -4.0625, | |
| "logits/rejected": -4.125, | |
| "logps/chosen": -130.0, | |
| "logps/rejected": -168.0, | |
| "loss": 0.5075, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -2.015625, | |
| "rewards/margins": 2.84375, | |
| "rewards/rejected": -4.84375, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.6932907348242812, | |
| "grad_norm": 12.343229925422747, | |
| "learning_rate": 2.420118343195266e-07, | |
| "logits/chosen": -4.09375, | |
| "logits/rejected": -4.09375, | |
| "logps/chosen": -143.0, | |
| "logps/rejected": -160.0, | |
| "loss": 0.5023, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -2.140625, | |
| "rewards/margins": 1.328125, | |
| "rewards/rejected": -3.484375, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.7252396166134185, | |
| "grad_norm": 12.271822104112585, | |
| "learning_rate": 2.36094674556213e-07, | |
| "logits/chosen": -4.09375, | |
| "logits/rejected": -4.125, | |
| "logps/chosen": -160.0, | |
| "logps/rejected": -185.0, | |
| "loss": 0.5209, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -3.421875, | |
| "rewards/margins": 1.609375, | |
| "rewards/rejected": -5.03125, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.7571884984025559, | |
| "grad_norm": 1.7194656245016546, | |
| "learning_rate": 2.301775147928994e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -4.03125, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -180.0, | |
| "loss": 0.4985, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -2.515625, | |
| "rewards/margins": 2.3125, | |
| "rewards/rejected": -4.8125, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.7891373801916934, | |
| "grad_norm": 9.434438571448306, | |
| "learning_rate": 2.242603550295858e-07, | |
| "logits/chosen": -4.0625, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -153.0, | |
| "logps/rejected": -206.0, | |
| "loss": 0.5254, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -3.234375, | |
| "rewards/margins": 2.53125, | |
| "rewards/rejected": -5.75, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.8210862619808306, | |
| "grad_norm": 3.770352466427391, | |
| "learning_rate": 2.1834319526627217e-07, | |
| "logits/chosen": -4.0, | |
| "logits/rejected": -4.09375, | |
| "logps/chosen": -140.0, | |
| "logps/rejected": -175.0, | |
| "loss": 0.5318, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -2.703125, | |
| "rewards/margins": 1.9765625, | |
| "rewards/rejected": -4.6875, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 1.8530351437699681, | |
| "grad_norm": 40.85212730947507, | |
| "learning_rate": 2.1242603550295858e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -4.0, | |
| "logps/chosen": -144.0, | |
| "logps/rejected": -154.0, | |
| "loss": 0.5218, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -2.984375, | |
| "rewards/margins": 1.53125, | |
| "rewards/rejected": -4.5, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 1.8849840255591053, | |
| "grad_norm": 9.19758791601065, | |
| "learning_rate": 2.0650887573964496e-07, | |
| "logits/chosen": -4.0625, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -176.0, | |
| "loss": 0.5163, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -3.046875, | |
| "rewards/margins": 2.1875, | |
| "rewards/rejected": -5.21875, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 1.9169329073482428, | |
| "grad_norm": 1.8606016727764927, | |
| "learning_rate": 2.0059171597633133e-07, | |
| "logits/chosen": -4.09375, | |
| "logits/rejected": -4.125, | |
| "logps/chosen": -154.0, | |
| "logps/rejected": -185.0, | |
| "loss": 0.5158, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -4.15625, | |
| "rewards/margins": 2.046875, | |
| "rewards/rejected": -6.21875, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.9488817891373802, | |
| "grad_norm": 6.5440298020464915, | |
| "learning_rate": 1.9467455621301774e-07, | |
| "logits/chosen": -4.09375, | |
| "logits/rejected": -4.09375, | |
| "logps/chosen": -156.0, | |
| "logps/rejected": -169.0, | |
| "loss": 0.5599, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -3.703125, | |
| "rewards/margins": 1.6484375, | |
| "rewards/rejected": -5.34375, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 1.9808306709265175, | |
| "grad_norm": 44.97604050874076, | |
| "learning_rate": 1.8875739644970412e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.953125, | |
| "logps/chosen": -156.0, | |
| "logps/rejected": -174.0, | |
| "loss": 0.5448, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -3.71875, | |
| "rewards/margins": 1.5625, | |
| "rewards/rejected": -5.28125, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -4.0, | |
| "eval_logits/rejected": -4.0, | |
| "eval_logps/chosen": -139.0, | |
| "eval_logps/rejected": -146.0, | |
| "eval_loss": 0.6721875071525574, | |
| "eval_rewards/accuracies": 0.1785714328289032, | |
| "eval_rewards/chosen": -3.5625, | |
| "eval_rewards/margins": 0.365234375, | |
| "eval_rewards/rejected": -3.9375, | |
| "eval_runtime": 12.0754, | |
| "eval_samples_per_second": 16.563, | |
| "eval_steps_per_second": 0.58, | |
| "step": 626 | |
| }, | |
| { | |
| "epoch": 2.012779552715655, | |
| "grad_norm": 5.387854507366258, | |
| "learning_rate": 1.8284023668639053e-07, | |
| "logits/chosen": -4.0625, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -161.0, | |
| "logps/rejected": -187.0, | |
| "loss": 0.5223, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -3.515625, | |
| "rewards/margins": 2.3125, | |
| "rewards/rejected": -5.8125, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 2.0447284345047922, | |
| "grad_norm": 1.2556332999951558, | |
| "learning_rate": 1.7692307692307693e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -147.0, | |
| "logps/rejected": -174.0, | |
| "loss": 0.4742, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -2.625, | |
| "rewards/margins": 2.703125, | |
| "rewards/rejected": -5.3125, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 2.07667731629393, | |
| "grad_norm": 1.0901290893409048, | |
| "learning_rate": 1.710059171597633e-07, | |
| "logits/chosen": -4.03125, | |
| "logits/rejected": -4.03125, | |
| "logps/chosen": -158.0, | |
| "logps/rejected": -181.0, | |
| "loss": 0.4933, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -3.734375, | |
| "rewards/margins": 2.5625, | |
| "rewards/rejected": -6.3125, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.108626198083067, | |
| "grad_norm": 0.7709648536735118, | |
| "learning_rate": 1.650887573964497e-07, | |
| "logits/chosen": -3.96875, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -145.0, | |
| "logps/rejected": -198.0, | |
| "loss": 0.4835, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": -2.984375, | |
| "rewards/margins": 4.15625, | |
| "rewards/rejected": -7.15625, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 2.1405750798722045, | |
| "grad_norm": 1.4684150448771174, | |
| "learning_rate": 1.591715976331361e-07, | |
| "logits/chosen": -3.96875, | |
| "logits/rejected": -4.03125, | |
| "logps/chosen": -157.0, | |
| "logps/rejected": -195.0, | |
| "loss": 0.4913, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -4.0625, | |
| "rewards/margins": 3.09375, | |
| "rewards/rejected": -7.125, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 2.1725239616613417, | |
| "grad_norm": 0.15445202520650775, | |
| "learning_rate": 1.5325443786982248e-07, | |
| "logits/chosen": -3.96875, | |
| "logits/rejected": -3.96875, | |
| "logps/chosen": -155.0, | |
| "logps/rejected": -190.0, | |
| "loss": 0.5065, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -3.734375, | |
| "rewards/margins": 2.84375, | |
| "rewards/rejected": -6.59375, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 2.2044728434504792, | |
| "grad_norm": 0.1682080177164145, | |
| "learning_rate": 1.4733727810650885e-07, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.921875, | |
| "logps/chosen": -161.0, | |
| "logps/rejected": -190.0, | |
| "loss": 0.5062, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 2.515625, | |
| "rewards/rejected": -6.90625, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.236421725239617, | |
| "grad_norm": 0.48752210028572696, | |
| "learning_rate": 1.4142011834319526e-07, | |
| "logits/chosen": -3.96875, | |
| "logits/rejected": -4.0, | |
| "logps/chosen": -167.0, | |
| "logps/rejected": -201.0, | |
| "loss": 0.4656, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -4.03125, | |
| "rewards/margins": 2.6875, | |
| "rewards/rejected": -6.71875, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 2.268370607028754, | |
| "grad_norm": 0.3925155000101438, | |
| "learning_rate": 1.3550295857988164e-07, | |
| "logits/chosen": -3.9375, | |
| "logits/rejected": -4.0, | |
| "logps/chosen": -168.0, | |
| "logps/rejected": -214.0, | |
| "loss": 0.4607, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 3.25, | |
| "rewards/rejected": -7.75, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 2.3003194888178915, | |
| "grad_norm": 0.4199125611952186, | |
| "learning_rate": 1.2958579881656802e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.96875, | |
| "logps/chosen": -172.0, | |
| "logps/rejected": -219.0, | |
| "loss": 0.504, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -4.46875, | |
| "rewards/margins": 3.53125, | |
| "rewards/rejected": -8.0, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 2.3322683706070286, | |
| "grad_norm": 0.18962721290307222, | |
| "learning_rate": 1.2366863905325443e-07, | |
| "logits/chosen": -3.921875, | |
| "logits/rejected": -4.03125, | |
| "logps/chosen": -167.0, | |
| "logps/rejected": -209.0, | |
| "loss": 0.4868, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -5.09375, | |
| "rewards/margins": 3.046875, | |
| "rewards/rejected": -8.125, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 2.364217252396166, | |
| "grad_norm": 0.15594202879866378, | |
| "learning_rate": 1.1775147928994082e-07, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -178.0, | |
| "logps/rejected": -198.0, | |
| "loss": 0.5083, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -5.21875, | |
| "rewards/margins": 2.546875, | |
| "rewards/rejected": -7.75, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 2.3961661341853033, | |
| "grad_norm": 2.995272322650428, | |
| "learning_rate": 1.1183431952662721e-07, | |
| "logits/chosen": -3.921875, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -181.0, | |
| "logps/rejected": -218.0, | |
| "loss": 0.4611, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": -5.09375, | |
| "rewards/margins": 2.984375, | |
| "rewards/rejected": -8.125, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 2.428115015974441, | |
| "grad_norm": 0.18498092374200545, | |
| "learning_rate": 1.059171597633136e-07, | |
| "logits/chosen": -3.890625, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -169.0, | |
| "logps/rejected": -186.0, | |
| "loss": 0.5076, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -5.28125, | |
| "rewards/margins": 1.953125, | |
| "rewards/rejected": -7.21875, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 2.460063897763578, | |
| "grad_norm": 0.08884257760247284, | |
| "learning_rate": 1e-07, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.921875, | |
| "logps/chosen": -145.0, | |
| "logps/rejected": -198.0, | |
| "loss": 0.4692, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 3.234375, | |
| "rewards/rejected": -7.53125, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 2.4920127795527156, | |
| "grad_norm": 2.04148740612421, | |
| "learning_rate": 9.408284023668639e-08, | |
| "logits/chosen": -3.921875, | |
| "logits/rejected": -3.953125, | |
| "logps/chosen": -158.0, | |
| "logps/rejected": -210.0, | |
| "loss": 0.4633, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": -3.90625, | |
| "rewards/margins": 4.21875, | |
| "rewards/rejected": -8.125, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 2.523961661341853, | |
| "grad_norm": 1.9100967302386418, | |
| "learning_rate": 8.816568047337278e-08, | |
| "logits/chosen": -3.984375, | |
| "logits/rejected": -4.0, | |
| "logps/chosen": -167.0, | |
| "logps/rejected": -200.0, | |
| "loss": 0.5019, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -5.125, | |
| "rewards/margins": 2.4375, | |
| "rewards/rejected": -7.5625, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 2.5559105431309903, | |
| "grad_norm": 0.42642836462978445, | |
| "learning_rate": 8.224852071005916e-08, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.921875, | |
| "logps/chosen": -184.0, | |
| "logps/rejected": -225.0, | |
| "loss": 0.4952, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -5.125, | |
| "rewards/margins": 3.078125, | |
| "rewards/rejected": -8.1875, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 2.587859424920128, | |
| "grad_norm": 0.1611701890949659, | |
| "learning_rate": 7.633136094674555e-08, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -168.0, | |
| "logps/rejected": -202.0, | |
| "loss": 0.4953, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -5.40625, | |
| "rewards/margins": 3.234375, | |
| "rewards/rejected": -8.625, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 2.619808306709265, | |
| "grad_norm": 0.15355437619982337, | |
| "learning_rate": 7.041420118343195e-08, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.953125, | |
| "logps/chosen": -182.0, | |
| "logps/rejected": -216.0, | |
| "loss": 0.502, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": -5.8125, | |
| "rewards/margins": 2.515625, | |
| "rewards/rejected": -8.3125, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 2.6517571884984026, | |
| "grad_norm": 0.23679653291952513, | |
| "learning_rate": 6.449704142011835e-08, | |
| "logits/chosen": -3.90625, | |
| "logits/rejected": -3.953125, | |
| "logps/chosen": -175.0, | |
| "logps/rejected": -210.0, | |
| "loss": 0.5132, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": -4.9375, | |
| "rewards/margins": 2.8125, | |
| "rewards/rejected": -7.75, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 2.68370607028754, | |
| "grad_norm": 1.1527288279228376, | |
| "learning_rate": 5.857988165680473e-08, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.859375, | |
| "logps/chosen": -166.0, | |
| "logps/rejected": -205.0, | |
| "loss": 0.4931, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -4.9375, | |
| "rewards/margins": 2.9375, | |
| "rewards/rejected": -7.875, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 2.7156549520766773, | |
| "grad_norm": 0.16029055506415082, | |
| "learning_rate": 5.266272189349112e-08, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -170.0, | |
| "logps/rejected": -222.0, | |
| "loss": 0.4838, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": -4.625, | |
| "rewards/margins": 4.0, | |
| "rewards/rejected": -8.625, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 2.747603833865815, | |
| "grad_norm": 0.20428026877833508, | |
| "learning_rate": 4.674556213017751e-08, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -175.0, | |
| "logps/rejected": -218.0, | |
| "loss": 0.4984, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -5.375, | |
| "rewards/margins": 3.21875, | |
| "rewards/rejected": -8.625, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 2.779552715654952, | |
| "grad_norm": 0.3728752675780405, | |
| "learning_rate": 4.082840236686391e-08, | |
| "logits/chosen": -3.953125, | |
| "logits/rejected": -3.953125, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -225.0, | |
| "loss": 0.4737, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -5.9375, | |
| "rewards/margins": 2.59375, | |
| "rewards/rejected": -8.5625, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 2.8115015974440896, | |
| "grad_norm": 0.20010578216519595, | |
| "learning_rate": 3.4911242603550294e-08, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.890625, | |
| "logps/chosen": -165.0, | |
| "logps/rejected": -185.0, | |
| "loss": 0.5236, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": -5.5625, | |
| "rewards/margins": 1.5625, | |
| "rewards/rejected": -7.125, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 2.8434504792332267, | |
| "grad_norm": 0.24869309029073208, | |
| "learning_rate": 2.8994082840236687e-08, | |
| "logits/chosen": -3.859375, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -180.0, | |
| "logps/rejected": -203.0, | |
| "loss": 0.4868, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": -5.0, | |
| "rewards/margins": 2.65625, | |
| "rewards/rejected": -7.65625, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 2.8753993610223643, | |
| "grad_norm": 0.1537738848665123, | |
| "learning_rate": 2.3076923076923076e-08, | |
| "logits/chosen": -3.84375, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -175.0, | |
| "logps/rejected": -224.0, | |
| "loss": 0.5018, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": -5.34375, | |
| "rewards/margins": 4.0, | |
| "rewards/rejected": -9.3125, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 2.9073482428115014, | |
| "grad_norm": 0.32035635694293296, | |
| "learning_rate": 1.7159763313609465e-08, | |
| "logits/chosen": -3.921875, | |
| "logits/rejected": -3.953125, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.5212, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": -6.03125, | |
| "rewards/margins": 2.421875, | |
| "rewards/rejected": -8.4375, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 2.939297124600639, | |
| "grad_norm": 0.407062337314559, | |
| "learning_rate": 1.1242603550295858e-08, | |
| "logits/chosen": -3.875, | |
| "logits/rejected": -3.875, | |
| "logps/chosen": -172.0, | |
| "logps/rejected": -202.0, | |
| "loss": 0.4832, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": -5.40625, | |
| "rewards/margins": 3.046875, | |
| "rewards/rejected": -8.4375, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 2.9712460063897765, | |
| "grad_norm": 0.20061865266174939, | |
| "learning_rate": 5.325443786982248e-09, | |
| "logits/chosen": -3.828125, | |
| "logits/rejected": -3.8125, | |
| "logps/chosen": -173.0, | |
| "logps/rejected": -205.0, | |
| "loss": 0.4932, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -5.125, | |
| "rewards/margins": 3.28125, | |
| "rewards/rejected": -8.4375, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -3.84375, | |
| "eval_logits/rejected": -3.859375, | |
| "eval_logps/chosen": -167.0, | |
| "eval_logps/rejected": -175.0, | |
| "eval_loss": 0.7137500047683716, | |
| "eval_rewards/accuracies": 0.1607142835855484, | |
| "eval_rewards/chosen": -6.375, | |
| "eval_rewards/margins": 0.53125, | |
| "eval_rewards/rejected": -6.90625, | |
| "eval_runtime": 14.9916, | |
| "eval_samples_per_second": 13.341, | |
| "eval_steps_per_second": 0.467, | |
| "step": 939 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 939, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5581134729111157, | |
| "train_runtime": 5075.7144, | |
| "train_samples_per_second": 5.91, | |
| "train_steps_per_second": 0.185 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 939, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |